Perception of synthetic emotion expressions in speech: Categorical and dimensional annotations

G. Bloothooft; J.M. Kessens; M.A. Neerincx; M. Kroes; R. Looije

doi:10.1109/ACII.2009.5349594

Perception of synthetic emotion expressions in speech: Categorical and dimensional annotations

G. Bloothooft, J.M. Kessens, M.A. Neerincx, M. Kroes, R. Looije

extern

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

Abstract

In this paper, both categorical and dimensional annotations have been made of neutral and emotional speech synthesis (anger, fear, sad, happy and relaxed). With various prosodic emotion manipulation techniques we found emotion classification rates of 40%, which is significantly above chance level (17%). The classification rates are higher for sentences that have a semantics matching the synthetic emotion. By manipulating the pitch and duration, differences in arousal were perceived whereas differences in valence were hardly perceived. Of the investigated emotion manipulation methods, EmoFilt and EmoSpeak performed very similar, except for the emotion fear. Copy synthesis did not perform well, probably caused by suboptimal alignments and the use of multiple speakers.

Original language	English
Title of host publication	3rd International Conference on Affective Computing and Intelligent Interaction and Workshops (ACII 2009)
Publisher	IEEE
Pages	303-307
Number of pages	5
ISBN (Electronic)	978-1-4244-4799-2
ISBN (Print)	978-1-4244-4800-5
DOIs	https://doi.org/10.1109/ACII.2009.5349594
Publication status	Published - 2009

Bibliographical note

Proceedings of a meeting held 10-12 September 2009, Amsterdam, Netherlands.

Access to Document

10.1109/ACII.2009.5349594

Cite this

@inproceedings{12a2a49422e442c39bcb0d04379ad186,

title = "Perception of synthetic emotion expressions in speech: Categorical and dimensional annotations",

abstract = "In this paper, both categorical and dimensional annotations have been made of neutral and emotional speech synthesis (anger, fear, sad, happy and relaxed). With various prosodic emotion manipulation techniques we found emotion classification rates of 40%, which is significantly above chance level (17%). The classification rates are higher for sentences that have a semantics matching the synthetic emotion. By manipulating the pitch and duration, differences in arousal were perceived whereas differences in valence were hardly perceived. Of the investigated emotion manipulation methods, EmoFilt and EmoSpeak performed very similar, except for the emotion fear. Copy synthesis did not perform well, probably caused by suboptimal alignments and the use of multiple speakers.",

author = "G. Bloothooft and J.M. Kessens and M.A. Neerincx and M. Kroes and R. Looije",

note = "Proceedings of a meeting held 10-12 September 2009, Amsterdam, Netherlands.",

year = "2009",

doi = "10.1109/ACII.2009.5349594",

language = "English",

isbn = "978-1-4244-4800-5",

pages = "303--307",

booktitle = "3rd International Conference on Affective Computing and Intelligent Interaction and Workshops (ACII 2009)",

publisher = "IEEE",

address = "United States",

}

Perception of synthetic emotion expressions in speech: Categorical and dimensional annotations. / Bloothooft, G.; Kessens, J.M.; Neerincx, M.A. et al.
3rd International Conference on Affective Computing and Intelligent Interaction and Workshops (ACII 2009). IEEE, 2009. p. 303-307.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Perception of synthetic emotion expressions in speech: Categorical and dimensional annotations

AU - Bloothooft, G.

AU - Kessens, J.M.

AU - Neerincx, M.A.

AU - Kroes, M.

AU - Looije, R.

N1 - Proceedings of a meeting held 10-12 September 2009, Amsterdam, Netherlands.

PY - 2009

Y1 - 2009

N2 - In this paper, both categorical and dimensional annotations have been made of neutral and emotional speech synthesis (anger, fear, sad, happy and relaxed). With various prosodic emotion manipulation techniques we found emotion classification rates of 40%, which is significantly above chance level (17%). The classification rates are higher for sentences that have a semantics matching the synthetic emotion. By manipulating the pitch and duration, differences in arousal were perceived whereas differences in valence were hardly perceived. Of the investigated emotion manipulation methods, EmoFilt and EmoSpeak performed very similar, except for the emotion fear. Copy synthesis did not perform well, probably caused by suboptimal alignments and the use of multiple speakers.

AB - In this paper, both categorical and dimensional annotations have been made of neutral and emotional speech synthesis (anger, fear, sad, happy and relaxed). With various prosodic emotion manipulation techniques we found emotion classification rates of 40%, which is significantly above chance level (17%). The classification rates are higher for sentences that have a semantics matching the synthetic emotion. By manipulating the pitch and duration, differences in arousal were perceived whereas differences in valence were hardly perceived. Of the investigated emotion manipulation methods, EmoFilt and EmoSpeak performed very similar, except for the emotion fear. Copy synthesis did not perform well, probably caused by suboptimal alignments and the use of multiple speakers.

U2 - 10.1109/ACII.2009.5349594

DO - 10.1109/ACII.2009.5349594

M3 - Conference contribution

SN - 978-1-4244-4800-5

SP - 303

EP - 307

BT - 3rd International Conference on Affective Computing and Intelligent Interaction and Workshops (ACII 2009)

PB - IEEE

ER -

Perception of synthetic emotion expressions in speech: Categorical and dimensional annotations

Abstract

Bibliographical note

Access to Document

Fingerprint

Cite this